A 3D biomimetic liver sinusoid construct for predicting physiology and toxicity Approximately 90% of drug candidates entering Phase 1 clinical trials fail, and one of the main reasons for drug failure is unexpected toxicity. The liver plays a centra role in the human body, contributing to homeostasis and important functions such as biotransformation and metabolism of drugs. The liver is also the most common target for drug-induced toxicity. Existing in vitro models and in vivo animal models have limited predictive power for human liver toxicity. The goal of this project is to construct a microfluidic liver modul which mimics the functions and responses of the human liver, with readouts designed to indicate both normal liver function and toxic responses. This human liver model is expected to be the essential elimination organ for modeling human exposure, provide improved predictions of drug induced liver toxicity, and also serve as a disease model for drug discovery. Our approach will be to develop a 3D microfluidic system with human hepatocyte, kupffer, stellate and endothelial cells, to mimic the liver acinus - the smallest functional unit of the liver. A uniue feature of the model will be the oxygenation of the media, and the establishment of an oxygen gradient, which is believed to account for important metabolic, gene expression and functional heterogeneity of the hepatocytes in the sinusoidal space of normal human liver. Hepatocytes in the oxygen rich zone are efficient in oxidative metabolism, fatty acid oxidation, gluconeogenesis, bile acid extraction, ammonia detoxification to urea and glutathione-conjugation while hepatocytes in the oxygen depleted zone are efficient in glycolysis, liponeogenesis and Cytochrome P-450 biotransformation. Another unique feature of the model will be the incorporation of 'sentinel' biosensor cells, a small fraction of cells with engineered biosensors that indicate changes in cellular functions. When combined with other fluorescent probes, standard biochemical and mass spectroscopy readouts, the model will provide a real-time High Content Analysis (HCA) profile to monitor organ function and response. The selection and validation of readouts and performance of the model will be evaluated based on a panel of reference drugs with available clinical data. To facilitate that comparison, a database of drugs with clinical data, and data from other in vitro and in vivo studies will be constructed. The ultimate goal of this project is to develop a microfluidic model of human liver function that will integrate with a series of other human organ modules, to create a microphysiology platform that reproduces human clinical trial results and provides improved predictivity of exposure, safety and efficacy for drug development. The liver plays a central role in human drug interactions, both within the liver and in other organs, as a result of drug metabolism. The performance of the liver module is central to the performance of the microphysiology platform. We believe the design proposed here will optimally recapitulate human liver function on that platform.